Bogdan Pasaniuc
@bpasaniuc
Followers
3K
Following
3K
Statuses
2K
Professor of Pathology&Lab Medicine, Human Genetics and Computational Medicine at UCLA. Views my own.
Los Angeles
Joined November 2012
We offer an individual-level interpretation of polygenic score (PGS) transferability across genetic ancestries: PGS accuracy varies individual-to-individual across the genetic-ancestry continuum, even in populations usually labeled as homogeneous. 1/
4
133
319
RT @ruthie_johnson: Very grateful to have worked on this project with such a fantastic group of people!
0
4
0
RT @PennGenetics: Excited to announce the Penn Postdoctoral Fellowship in Genetics! We seek to empower and nurture talented early-stage res…
0
25
0
RT @mikejg84: Very excited to see isoTWAS published today in @NatureGenet! Big power gains for complex trait mapping via multivariate mo…
0
27
0
RT @hjpimentel: Welcome back! If you're looking for a faculty position in the intersection of computational methods and medicine [broadly],…
0
33
0
polygenic👏score👏means👏cannot👏be👏compared👏across👏populations [Figure modified from Ding et al.
0
0
5
@nmancuso_ I interpret this type of comments as “not focused enough” on what deliverables will be provided at end of the research plan. A good review should be explicit on what is “ambitious” and what is not.
1
0
3
RT @GENES_PK: @bpasaniuc @nmancuso_ @anshulkundaje @michelnivard @kangchenghou @genetisaur You may ask yourself, what is this Connect thing…
0
1
0
@nmancuso_ @anshulkundaje @michelnivard @kangchenghou @genetisaur Critical to consider is the people that consented to be part of AoU themselves. Do we know what they consented for with respects to data sharing? Do participation rates vary across different data sharing architectures? How is data sharing considered for future data @GENES_PK ?
1
0
0
@nmancuso_ @hjpimentel @anshulkundaje @kangchenghou @genetisaur Great point! And that speaks towards not forcing only one option but allowing for either cloud or local depending on PI/project/etc.
0
0
1
@nmancuso_ @hjpimentel @anshulkundaje @kangchenghou @genetisaur big bonus if by taking over compliance/security NIH/AoU drastically reduces the paperwork and number of signatures on all sorts of docs.
2
0
1
@hjpimentel @anshulkundaje @nmancuso_ @kangchenghou @genetisaur Yes. The counter is wait till you have to explain to the data security signing official for the dbgap application that human genetic data will sit in this corner office :)
2
0
2
@anshulkundaje @nmancuso_ @kangchenghou @genetisaur Yep; we’re all in violent agreement! I do think the community can come together around a set of principles that should be respected by the cloud infrastructure. Here is our pro/con experience for all of us cloud; curious to learn about others:
@genetisaur @GENES_PK @nmancuso_ @kangchenghou ditto! from our exp, pros: ease of access; little PI gatekeeping; sandbox to test out new ideas outside clearly scoped projects; super easy to set up collaborations thru shared workspaces; straightforward system to get manuscripts approved. 1/2
0
1
0
@anshulkundaje @nmancuso_ @kangchenghou @genetisaur I am often reminded that smaller groups/institutions may find it hard to justify heavy compute infrastructure investments for one-off research projects on these large-scale biobank datasets. So all in all, it’s really complicated with no single solution.
1
0
0
@bpasaniuc @nmancuso_ @kangchenghou @genetisaur Would love to hear specific pros and cons of All of Us etc, things to copy or avoid. Thinking of NCI Connect cohort—we’re in the weeds of recruiting and data collection now but very soon will need a way to share these data with research community.
0
0
1